The genetic dissection of complex traits

The genetic dissection
of complex traits
Karl W Broman
Biostatistics & Medical Informatics
University of Wisconsin – Madison
www.biostat.wisc.edu/~kbroman

The goal
Identify genes that contribute to complex diseases
Complex disease = one that’s hard to figure out
Many genes + environment + other stuff
2

The genetic approach
• Start with the trait; find genes that influence it
– Allelic differences at the gene result in phenotypic differences
• Value: Need not know anything in advance
• Goal
– Understand the disease etiology (pathways/mechanisms)
– Prediction/prevention
– Identify possible drug targets
3

Approaches
• Experimental crosses in model organisms
• Mutagenesis in model organisms
• Association analysis with inbred strains
• Linkage analysis in human pedigrees
– A few large pedigrees
– Many small families (e.g., sib pairs)
• Association analysis in human populations
– Candidate genes vs. whole genome
4

Human vs mouse
www.daviddeen.com
6

Mutagenesis
Advantages
+ Can find things
+ At the gene
Disadvantages
– Need cheap phenotype screen
– Mutations must have large effect
– Genes found may be irrelevant
– Still need to map the mutation
– Recessive mutations are hard to see
7

Intercross
P 1 P 2
F 1 F 1
F 2
8

LOD curves
8
6
LOD score
4
5%
2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Chromosome
9

Joys of QTL mapping
• Genotype → phenotype
• Recombination is cool
• Simple correlation structure
• Lots of opportunity for collaboration
• Lots of open problems
• Not many competitors
10

Chromosome 4
8
6
LOD score
4
2
0
1.5−LOD support interval
0 20 40 60
Map position (cM)
11

Traditional approach
F 2 /BC −→ QTL −→ congenic −→ subcongenics
−→ • coding/regulatory polymorphism
• expression/function difference
• knock-in / transgenic
• knock-out
• homology to other species
Issues:
• Large QTL regions
• Time consuming and expensive
12

A congenic line
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 X Y
13

More modern approaches
• Gene expression data
– and proteins, metabolites, epigenetic marks
– and eQTL analyses
• Consomics (aka chromosome substitution strains)
• Panels of congenics
• Multiple crosses + haplotype analysis
• Cross-species comparisons
• Advanced intercross lines / Heterogenous Stock
• Recombinant inbred lines / Collaborative Cross
• Association mapping with inbred lines
14

Consomics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 X Y 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 X Y 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 X Y
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 X Y 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 X Y 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 X Y
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 X Y 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 X Y 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 X Y
15

Consomics
Advantages
+ Just phenotyping can get you to
the chromosomes
+ Eliminate the effects of other QTL
+ Easy to create congenics
Disadvantages
– Time-consuming, expensive to
create
– Lots of phenotyping required
– Cannot see interactions
16

B.A on low fat diet
C57BL/6J − Chr A J / NaJ
0.10
0.08
Weight gain
0.06
0.04
0.02
0.00
A/J 7 10 14 11 6 1 Y 9 4 16 8 X 12 2 19 3 17 13 15 M 5 18
B6
Strain
Shao et al., PNAS, 105:19910–19914, 2008
17

More modern approaches
• Gene expression data
– and proteins, metabolites, epigenetic marks
– and eQTL analyses
• Consomics (aka chromosome substitution strains)
• Panels of congenics
• Multiple crosses + haplotype analysis
• Cross-species comparisons
• Advanced intercross lines / Heterogenous Stock
• Recombinant inbred lines / Collaborative Cross
• Association mapping with inbred lines
18

Advanced intercross lines
P
A
B
F 2
F 3
F 4
F 7
F 10
19

AIL
Advantages
+ Many more breakpoints =⇒
more precise mapping
+ Straightforward to create
Disadvantages
– Time and cost
– Each individual genetically distinct
– Useful largely for fine-mapping
known loci
– Relationships among individuals
in final generation
20

Sibships at F 8
21

AIL: percent matching
genotypes
22

AIL: genotype data
23

Heterogeneous stock
G 0
A B C D E F G H
G 1
G 2
G 10
G 20
G 40
24

HS
Advantages
+ Super-dense breakpoints
+ Many alleles
+ Heterozygous
Disadvantages
– Must be satisfied with what is
available
– Inbreeding: loss of alleles
– Each individual unique
– Like AIL, maybe best for
fine-mapping known loci
– Like AIL, relationships at last
generations
25

Recombinant inbred lines
P 1 P 2
F 1 F 1
F 2
F 3
F 4
●
●
●
F ∞
26

RIL
Advantages
+ High density of breakpoints
+ Just genotype once
+ Phenotype multiple individuals to
reduce environmental/individual
variation
+ Multiple phenotypes on the same
genomes
+ Longitudinal phenotypes
+ Genotype × environment
interactions
Disadvantages
– Time-consuming, expensive to
create
– Available panels generally too
small
– Only homozygotes
27

RIX
1 2 3 4 5
RI strains
RIX
1 x 2 1 x 3 1 x 4 1 x 5 2 x 3 2 x 4 2 x 5 3 x 4 3 x 5 4 x 5
28

Collaborative Cross
G 0
A B C D E F G H
G 1
A B C D E F G H
G 2
ABCD EFGH
G 3
G 4
●
●
●
G ∞
29

A CC genome
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 X Y
30

Association mapping
• Phenotype available inbred strains
• Make use of available SNP data
• Need to account for the correlations among strains
• Likely want to work with haplotypes rather than just individual SNPs
• Be careful about wild-derived strains
31

Association mapping
Advantages
+ Once you’ve done a strain survey,
no further data needed
+ Potentially very high resolution
Disadvantages
– All the usual problems with
association mapping
– Power is unpredictable
– How to account for relationships
among strains
32

CC vs HS vs
association mapping
These approaches have many similarites.
Key differences:
• CC, HS: pattern of association along chromosomes by design
• HS: each individual unique
33

Summary
• QTL mapping in mice is fun and useful
• But we need to be able to get to gene
• There are lots of new strategies to speed things along
• Numerous interesting statistical problems remain
• Need to consider human GWAS
34